.jpg)
Snowflake has announced a suite of new AI innovations designed to build a trusted, scalable foundation for enterprise artificial intelligence. Central to this is Semantic View Autopilot, an AI-powered service that automates the creation and governance of business logic, ensuring AI agents operate on consistent, accurate metrics. The company also unveiled significant advancements for machine learning development and tools to evaluate and audit the behavior of production AI agents.
Snowflake announces Semantic View Autopilot (GA) to automate the creation and governance of semantic business views.
The goal is to give AI agents a shared, trusted understanding of business metrics to prevent errors and hallucinations.
Snowflake Notebooks (GA) integrates with Cortex Code for building ML pipelines via natural language prompts.
New Cortex Agent Evaluations (coming soon) provide visibility into agent reasoning for auditing and improvement.
Enhanced Cortex AI Functions include cost governance tools to estimate and control AI spending.
Early adopters include eSentire, HiBob, Simon AI, VTS, Aimpoint Digital, and WHOOP.
A core challenge for enterprise AI is the inconsistent, manual definition of business metrics across different tools, which leads to unreliable AI outputs. Semantic View Autopilot addresses this by using AI to automatically build, optimize, and maintain governed semantic views. This creates a single source of truth for business logic that can be shared with AI agents and consumption tools like dbt Labs, Looker, Sigma, and ThoughtSpot. The service aims to cut semantic model creation from days to minutes while ensuring AI agents deliver accurate, consistent results based on trusted business definitions.
To streamline the development of powerful ML models, Snowflake has made Snowflake Notebooks generally available. This Jupyter-powered notebook is integrated with the Cortex Code AI agent, allowing data scientists to build and deploy complete ML pipelines using simple natural language prompts. For production, new features like Online Feature Store and Online Model Inference enable real-time predictions. Experiment Tracking within the notebooks helps teams compare training runs and reproduce the best models, turning experimentation into a repeatable process.
As AI agents take on mission-critical tasks, ensuring their behavior is trustworthy and auditable becomes paramount. Cortex Agent Evaluations (coming soon) provide developers with deep visibility into how agents reason and act, enabling systematic assessment of answer correctness and logical consistency. This allows teams to identify errors, refine logic, and validate performance before deployment. Furthermore, Snowflake is expanding Cortex AI Functions with cost governance capabilities, such as the AI_COUNT_TOKENS function, to help organizations proactively estimate and control AI spending.
Snowflake's latest announcements represent a comprehensive push to solve the operational and governance challenges that hinder enterprise AI at scale. By automating the foundational business logic layer, accelerating the ML lifecycle, and providing tools for agent evaluation and cost control, Snowflake is positioning its platform not just as a place to run AI models, but as an integrated environment where AI can be built, trusted, governed, and scaled with confidence. This focus on the entire AI value chain—from trusted data to measurable outcomes—is key to transitioning AI from experimental projects to core components of business operations.
About Snowflake
Snowflake is the platform for the AI era, making it easy for enterprises to innovate faster and get more value from data. More than 12,600 customers around the globe, including hundreds of the world’s largest companies, use Snowflake’s AI Data Cloud to build, use and share data, applications and AI. With Snowflake, data and AI are transformative for everyone.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.